|
Title
|
|
|
|
Utilizing deep learning and advanced image processing techniques to investigate the microstructure of a waxy bitumen
|
|
|
Author
|
|
|
|
|
|
|
Abstract
|
|
|
|
| Bitumen, also called asphalt binder, is the key component in asphalt mixtures. Studies to investigate the microstructure of this material show a rich morphology, especially the formation of bee structures in bitumen containing wax. Most research in this field has investigated these microstructures using commercial image processing software that needs a manual selection of these patterns to obtain certain characteristics. This study aims to construct a deep-learning-based object-detection model that can detect these bee patterns in the images acquired from bitumen samples using Confocal Laser Scanning Microscopy (CLSM). The CLSM images are then used to determine the morphological properties of the samples. The properties investigated are some typical roughness parameters and the wavelength calculated by a novel image processing technique based on the two-dimensional fast Fourier transform. In addition, these developed methodologies are used to investigate the influence of short-term and long-term ageing on the microstructure of a waxy bitumen. The results show that the trained deep learning model can be used to successfully detect the location, number, and area of the bee structures. The number of bee patterns and the area of the surface they cover are reduced upon ageing. Furthermore, some strong trends are found between the computed roughness parameters and the ageing level of the samples. Finally, the estimated wavelength of the bee patterns increases by ageing bitumen. The successful development and demonstration of these methods show their great potential in analyzing the microscopic images of bitumen taken by CLSM or atomic force microscopy and the enormous opportunities for future research. |
|
|
|
Language
|
|
|
|
English
|
|
|
Source (journal)
|
|
|
|
Construction and building materials. - Reigate
|
|
|
Publication
|
|
|
|
Reigate
:
2021
|
|
|
ISSN
|
|
|
|
0950-0618
|
|
|
DOI
|
|
|
|
10.1016/J.CONBUILDMAT.2021.125481
|
|
|
Volume/pages
|
|
|
|
313
(2021)
, 9 p.
|
|
|
ISI
|
|
|
|
000718499400001
|
|
|
Full text (Publisher's DOI)
|
|
|
|
|
|
|
Full text (open access)
|
|
|
|
|
|
|
Full text (publisher's version - intranet only)
|
|
|
|
|
|